I'm thinking of buying this Admiral

[dobulee]

Don't get discouraged! I don't know what your background with regard to electronics is, but you've made this set look pretty good in my opinion.

Even as a practicing engineer, I look at a lot of 40's schematics and say either 'what the heck is that', or 'how in the world did they come up with THAT!'

So, interestingly enough, the 1950 Breitman manual shows a ton of schematics, more than half of them have DC restoration. (interestingly none of the cathode driven sets seem to have it, but I think thats coincidence) The 1951 Breitman manual shows the next years sets, and only a handful have DC restoration. Even sets from GE lack it. I'm going to say that with content of the era, average video brightness didn't change much, and manufacturers found they could save one tube with minimal impact.

If you're interested, we can talk more about DC restoration either offline (or in the forum). It's not unique to TV. The basic circuit only takes 2 or 3 parts (1 diode, 1 capacitor, 1 resistor). Its used any time that you have to AC couple a signal for amplification or other reasons, but then want to 'restore' the DC information that was stripped off the signal when you AC coupled it.

As you know, the average video voltage is low for an all black screen, and high for an all white screen. The DC information is ALWAYs lost in broadcast reception, and even if it weren't it would be lost in the amplification stages. (and actually for your set, it's reversed since you have cathode drive which reverses the polarities) If you fed your set an all black screen, it would eventually settle in to '50% brightness). If you then immediately fed it an all white image, it would flash white then fade to 50% brightness again once the DC image level ceased to pass the amplifier.
TVs use the level of the synch tips as the DC reference level for the picture - however the absolute voltage of the synch tips also changes over time with picture brightness. So, we create a circuit that looks at the 'local' height of the synch tips and recreates the picture information relative to that. the time constant of the circuit allows our reference level to change over time as brightness level changes.

I have no experience picking the time constant required, but we can look at other sets for ideas. What I've attached seems typical.

I came up with the attached schematic. and ran a quick simulation with some video waveforms. I *think* it should work well enough. Though, you'll probably have to tweak the values. The diode can be any low capacitance signal/detector diode. You'd be wise to go with 400V or greater. Play with R5 some - my guess is that small values (or zero) will make dc restoration better but may cut down video bandwidth. larger values will not restore dc as well, but will have less effect on video. This circuit also adds an additional cap into the video chain - it may or may not have an effect on video bandwidth. try it and see. Once you find a place for the circuit, you can play with the values.

I attached a picture of my 1948 RCA 6T45 test pattern, as well as a live image. I dare say that you have a much better picture.

[dobulee]

So...hopefully you didn't dig into that circuit too much

It should have been obvious, but that circuit will also lose the DC level over time.

Cathode driven DC restoration is indeed a little bit more difficult since the set requries the cathode voltage to be about midway between the video amp plate and GND.

Anyhow, I gave it some more thought. We need a way to clamp the DC level to a reasonable voltage that is stiff. What I came up with uses the same 2 470k resistors to set the cathode bias, but rather than come off the video amp output, the 'top' 470k connects to C40 -- the same place the brightness pot taps off of. Those resistors give you a voltage at approximately half the video amp screen voltage, which is then stiffened by a 5uF (or so) 250V or better electrolytic.

D1 clamps the black level relative to the new 250V point. C2 and R6 set the time constant of the DC restoration (it's set similar to other sets of this era). The comments about R5 still stand - bigger R5 gives worse DC restoration, but might give better video and/or avoid synch problems. Give it s spin and see what works.

I've attached the new schematic.

I attached some simulated waveforms of what you might see at the original cathode and at the cathode with this circuit. The DC restoration isn't perfect, but it should be better than what you have.

I created some simulated composite video with positive synch (like you should have at the output of your video amp). The simulation has a bunch of lines of high brightness followed by a completely black section of video, where there is nothing above the synch pulses.

The original composite video is at top, the DC restored cathode signal in the middle, and the original circuit's cathode at bottom.

I'm not sure the AC coupled composite simulation is entirely accurate...but at least it shows how the set's original video amp causes black to fade to gray, while the new circuit eventually finds the right black level after an abrupt brightness change.

Nothing's perfect, hopefully this is something that others can common on and improve upon. After looking I was unable to find any actual schematics of cathode driven sets with DC restoration.